Dontopedia

Standard Backprop

From Dontopedia, the open, paraconsistent wiki. (Last updated 2026-06-06.)

Standard Backprop has 8 facts recorded in Dontopedia across 2 references, with 1 live disagreement.

8 facts·6 predicates·2 sources·1 in dispute

Mostly:step includes(3), requires backward pass(1), step1(1)

Maturity scale raw canonical shape-checked rule-derived certified

Inbound mentions (3)

Other subjects in dontopedia point AT this entity as a value. These are inverse relationships — e.g. "X motherOf this subject" — and answer questions the forward facts can't. Grouped by predicate.

usesUses(2)

usesStrategyUses Strategy(1)

Other facts (8)

The long tail: predicates that appear too rarely to warrant their own section. Filter or scroll to find a specific one. Each row links to its source.

8 facts
PredicateValueRef
Step IncludesForward Pass[2]
Step IncludesBackward Pass[2]
Step IncludesGradient Update[2]
Requires Backward Passtrue[1]
Step1Forward pass → compute loss[1]
Step2nn.value_and_grad() → autodiff backward pass through every layer[1]
Step3Gradients tell each parameter exactly which direction reduces loss[1]
Step4Update parameters accordingly[1]

Timeline

Timeline axis is valid_time — when each source says the fact was true in the world, not when Dontopedia learned about it. Retracted rows are kept for provenance; coloured stripes indicate the context kind.

requiresBackwardPassblah/watt-activation/part-122
true
step1blah/watt-activation/part-122
Forward pass → compute loss
step2blah/watt-activation/part-122
nn.value_and_grad() → autodiff backward pass through every layer
step3blah/watt-activation/part-122
Gradients tell each parameter exactly which direction reduces loss
step4blah/watt-activation/part-122
Update parameters accordingly
stepIncludesblah/watt-activation/122
ex:forward-pass
stepIncludesblah/watt-activation/122
ex:backward-pass
stepIncludesblah/watt-activation/122
ex:gradient-update

References (2)

2 references
  1. [1]Part 1225 facts
    ctx:discord/blah/watt-activation/part-122
  2. [2]1223 facts
    ctx:discord/blah/watt-activation/122
    • full textwatt-activation-122
      text/plain3 KBdoc:agent/watt-activation-122/57649dd0-cec5-4d9a-bc09-bec5f2db2137
      Show excerpt
      [2026-03-09 01:19] xenonfun: ⏺ BP = Backpropagation — whether the optimizer computes gradients via automatic differentiation or not. Adam / RotAdamW use standard backprop: 1. Forward pass → compute loss 2. nn.value_and_grad() → autod

See also

Keep researching

Missing something or suspicious of what's here? Kick off a research session — a Claude agent will investigate, cite its sources, and file new facts into a dedicated context you can review before accepting into the shared view.